Meta is rolling out AI-powered age-estimation technology globally this month, starting with Instagram and extending to Facebook and Threads later in 2025. The system analyzes user behavior and posts to identify accounts likely belonging to teenagers, even when those users have entered a false adult birthday. Accounts flagged as under-18 are automatically placed into Teen Account protections — restricted contacts, content filtering, usage caps, and nighttime notification curfews. The timing is not coincidental. In August 2024, Meta agreed to an $18 billion settlement with US attorneys general who alleged the company's platforms encouraged addictive behavior, failed to verify ages, and amplified harmful content targeting minors. The settlement mandated specific changes: two-hour daily usage limits, midnight-to-6am curfews, suppression of social-comparison features like visible likes, and parental notification systems. What Meta now presents as voluntary global expansion is, in significant part, the fulfillment of court-ordered obligations dressed as corporate initiative. The restrictions themselves are substantial. Teen users face push-notification blackouts during school hours (8am-3pm), break prompts every 15 minutes, and alerts at 60 and 90 minutes of continuous use. Searches for suicide, self-harm, or eating disorders trigger parental notifications. Cosmetic-procedure filters that alter facial appearance are blocked by default. Parents and guardians receive activity updates and first-contact alerts when teens message adults. These are real structural interventions, not cosmetic changes. The age-estimation technology itself raises the central tension. To detect teens who lie about their age, Meta must conduct behavioral profiling of every user — analyzing posting patterns, content consumption, writing style, and interaction habits to infer demographic characteristics. This is surveillance infrastructure with a sympathetic use case. The same system that identifies a 14-year-old pretending to be 18 also builds granular behavioral models of every adult on the platform. Meta controls what the system detects, how it classifies, and what it does with the underlying data. Australia's approach — an outright ban on social media for under-16s — is already facing legal challenge. Reddit and the Digital Freedom Project have filed suit questioning both the constitutionality and the scientific rationale of total platform exclusion. Meta's age-estimation model occupies the middle ground: allow teens on the platform but algorithmically restrict their experience. This positions Meta as the reasonable actor compared to outright bans, while consolidating its role as the arbiter of what content is appropriate for which users. Meta's call for TikTok and YouTube to adopt identical practices is strategic, not altruistic. If age-estimation becomes an industry standard, the companies with the largest behavioral datasets and the most advanced AI infrastructure gain a structural advantage. Smaller platforms and new entrants face compliance costs that function as barriers to entry. The child-safety mandate becomes, over time, a competitive moat. The fundamental question is whether a company that built its business on maximizing teen engagement — and paid $18 billion for the consequences — should also be the company that designs and operates the surveillance apparatus meant to limit that engagement. The fox is building the henhouse security system, and the blueprints are proprietary.
